By the rivers of PCT

[From Rick Marken (930618.1900)]

Allan Randall (930618.1800 EDT) --

I said:

You can use any kind of control system you like; or any method
you like whatsoever -- it doesn't even have to be a control system.

Allan says:

No, I need to know what control system to do the reconstruction for,
not what to use to do the reconstruction.

I gave you, along with the disturbance, EVERY parameter of the control
system except for the function relating output to perception (the
feedback function). Surely you don't consider the feedback function
to be part of the control system, do you? In a tracking task that is
the function that relates my muscle tensions (the last output of my
nervous system) to the cursor position. Surely you don't consider my
desk, the rollers in my mouse, the electronic characteristics of
the computer, etc to be part of me as a control system, do you?

Correct me if I'm
misremembering: You have stated that the simple linear example from
Powers' Primer was insufficient due to lack of nonlinearities.

You were able to reconstruct the disturbance when you knew p, r
and O(); ie. when I gave you every component of the control
system. Then I sent you p along with r and O() but with a feedback
function that was not = 1 (and I didn't tell you about the feedback
function; but that's not part of the control system so why bother).
I thought you would try the reconstruction and find that your
computed disturbance was not equal to the actual one anymore.
But you didn't do it, so you never found out. You didn't try the
reconstruction, you say, because:

this is an
EMPTY challenge. It has nothing whatsoever to do with anything I have
claimed about information in the percept or reconstructibility. You are
asking for a completely 100% blind reconstruction. This would require 100%
of the information about the disturbance in the percept. I never claimed

So fine. So there is not enough information in perception from
which to reconstruct the disturbance. Great. We are then agreed,
except perhaps on the details. You say that the amount of information
in perception about the disturbance is >0; I say that it is precisely
0. So all this was wasted effort. Clearly, the control system cannot
use the information in perception to produce effects that exactly
counter the disturbance because there is NOT ENOUGH information to do
it. OK. You can believe that there's a little weensey amount of
information in the perceptual signal if you like. Great. What's it
good for? What does the control system "use" to produce effects on
the controlled variable that are exactly opposed to the disturbance?

Or are you saying that your reconstruction would be only 99% accurate?
That would be GREAT. I'd accept a reconstruction as close to the
disturbance as was the output of the control system. In fact, why not
just reconstruct the output -- forget the disturbance. I'll again give
you the perception, reference and output function. Surely there is
enough information in the perception from which to reconstruct the

I think it was you, Allan, who said that the input to the control system
must have enough information so that it can tell the output of the
control system what to do to counter the disturbance. That's why I gave
you the reference and output function for your reconstruction -- your
concept of control seemed to be:

p --> (r-p) --> i -->O()--> o -->CV --> p

The perception comes into the system, is subtracted from a reference
and the information (i) is extracted. This information is used to
tell the "plant" (the output function, O()) what to do. The plant
produces an output (o) that will counter the effect of d,
resulting in a p closer to r. I believe you came to the conclusion that
to believe that control did not work something like the way I diagrammed
it above was to believe in magic (or mysticism). Martin said that to believe
otherwise would be a "renunciation of logic".

You wanna try reconstructing the output from the perception, reference
and output function. Is there at least enough information to do that???

The fact is that you guys just have a wrong conception of how control
works -- its a circle, not a chain. Your belief in this information
stuff is keeping you from making the small but profound mental
change that is needed to be able to visualize how a control system
actually works. I'm telling you, you are missing a truly profound
experience by holding on to the life raft of information. Let go,
and swim through the refreshing waters of control of perception
(if you want to, of course).



[Martin Taylor 930621 11:00]
(Rick Marken 930618.1900)

A *slight* inconsistency here, quoted without comment:


Surely you don't consider the feedback function
to be part of the control system, do you?


The fact is that you guys just have a wrong conception of how control
works -- its a circle, not a chain.

Do you see the contradiction? We keep trying to tell you that we NEED the
circle, and you keep trying to tie us up with the chain. You deny us in
your challenges the opportunity to use the necessary circle, and then
tell us our analysis won't work because we think we don't need the



I think it was in a very early posting on this information theory bit that
I pointed out that the fact of control demanded that the uncertainty
about the world remained stable over time. That, in iteself, demonstrated
the need for the completed feedback circuit. It would be nice if you could
get OUT of your head the notion that an information-theoretic approach
deals with an open-ended system.